Geography Reference
In-Depth Information
spectral and temporal resolution) provide opportunities
to progress quickly in this domain and significant devel-
opments are expected in the forthcoming years.
For network-scale managers, until the 1990s, the practi-
cal choice was between broad-scene but low resolution or
narrow-scene and good spatial resolution. In the first case,
when riparian vegetation is a narrow strip along the river,
an important part of the network can not be mapped.
In the second, due to the form and spatial distribution
of riparian vegetation, the number of images needed
for broad-scale characterisation is problematic in terms
of data treatment. Advances in classification procedures,
increased spatial resolution, GIS integration and comput-
ing capacities all contribute towards increased accuracy in
reach scale vegetation mapping. Furthermore, these are
also used increasingly in integrated mapping studies that
combine broad-scale and detailed vegetation type char-
acterisation (Claggett et al., 2010; Wiederkher et al., 2010;
Tormos et al., 2010; Johansen et al., 2010c). For example,
with a combination of high spatial resolution images
(QuickBird) and an object-oriented procedure, Johansen
et al., (2007b) discriminated seven different vegetation
types within a 10 km by 5 km area. There are some specific
limitations at such scale, for example homogeneity, infor-
mation extraction or cost of images. Some other limita-
tions are the same as for the reach scale, but the objectives
usually differ: it is a general but exhaustive description
that is reached rather than a detailed one. For this pur-
pose, images have good potential. In conclusion, image
analysis can provide some complementary sets of data to
combine with field information (Brooks et al., 2009) and
objectives should be clearly identified before application.
vegetation studies, ranging from a simple and accessible
application using aerial photography to more sophisti-
cated ones using LiDAR and Radar data.
10.5.1 Alow-cost strategy formonitoring
changes inafloodplainforest: aerial
photographs
A river restoration project, supported by the European
community LIFE program, has been launched in order to
manage and restore ecological, landscape and recreational
elements of the remaining alluvial forest patches within
the Arve River valley. The Arve River is an intra-mountain
tributary of the Rh one River flowing from Mont Blanc
(French Alps). Until the beginning of the twentieth cen-
tury, the Arve was a braided river system, except locally
where embankments had been previously constructed. In
the aftermath of the Second World War, intensive gravel
extraction resulted in active channel degradation, corri-
dor fragmentation and subsequent changes of community
composition in the floodplain forest. Remaining alluvial
forest patches consisted of approximately 600 ha of wood-
land along a 50 km reach. The restoration project was led
by the French Forestry commission (ONF), local districts,
and river managers with scientific support of the CNRS,
the French national organisation for scientific research.
While carrying out the project, managers were confronted
with common limitations including a lack of historical
surveys, limited funding for field work and a complex
mosaic of communities. As a result, historical imagery was
employed as the primary tool to describe the temporal
changes of the landscape and riparian vegetation.
10.5.1.1 Landscape changes: historical changes
of the vegetation within the corridor
10.5 Examples of imagery applications
and potentials for riparian
vegetation study
Between 1936 and 2004, 11 different aerial surveys were
carried out over the Arve River. Of these, five dates were
selected for comparison so as to limit the cost of the analy-
sis. The choice was made (1) to select the earliest and latest
dates (1936 and 2000), (2) to reject poor-quality and low-
resolution images, and (3) to select dates with good tem-
poral resolution (i.e., decadal sampling in 1961, 1973 and
1984) during the most dynamic periods (Greco and Plant,
2003). All of the photographs have been rectified and geo-
referenced using ArcGIS software, and spatial scales varied
between 1/15 000 and 1/25 000. The photo series were
black and white until 1973, infrared colour in 1984 and in
colour in 2000. For each date, a map of the land cover was
drawn by photo interpretation with classes defined by the
active river zone (unvegetated bars and main channel),
Documenting and analysing the structure and changes of
riparian vegetation is an important step for river man-
agers, forest services and conservation biologists involved
in conservation and restoration projects (Van Looy et al.,
2008; Brierley and Fryirs, 2008; Dufour and Piegay, 2009).
It is also a fundamental point for hydraulic engineers that
have to set up flood control plans and thus evaluate river
and floodplain roughness (Kouwen, 2000; Straatsma and
Middelkoop, 2007). Lastly, for network-scale managers,
riparian buffers mapping is crucial to assess physical and
chemical river quality (Claggett et al., 2010). Thus, we
now present three examples of imagery use in riparian
Search WWH ::
Custom Search